← Authors

Kefan Dong

13 papers · 2019–2025 · 5 conferences · across top CS/AI conferences

Achievements

🧭 Keyword Pioneer 🐣 Hot Topic Early Bird 🌈 Renaissance Researcher (5) πŸŒ‰ Interdisciplinary Bridge πŸ—ΊοΈ Taxonomy Completionist (18) 🌍 Conference Polyglot (5) πŸƒ Academic Marathon (6) 🐝 Cross-Pollinator (13) πŸ† Grand Slam πŸ’Ž Century Club (13) ⚑ Prolific Year (5)

Conferences

NIPS (4) ICLR (3) ICML (3) COLT (2) AAAI (1)

Top co-authors

Tengyu Ma (7) Yuan Zhou (3) Jian Peng (2) Emma Brunskill (2) Arvind Mahankali (1) Jiaqi Yang (1) Zhizhou Ren (1) Yuping Luo (1) Xiaoyu Chen (1) Yining Wang (1)

Keywords

neural network (3) contextual bandit (2) model-based reinforcement learning (2) sample complexity (2) online learning (2) policy optimization (1) markov decision process (1) learning theory (1) model-based learning (1) function approximation (1) offline reinforcement learning (1) model misspecification (1) gradient descent (1) policy selection (1) batch learning (1) robotic manipulation (1) spectral analysis (1) assortment optimization (1) ucb algorithm (1) reinforcement learning (1)

Papers

STP: Self-play LLM Theorem Provers with Iterative Conjecturing and Proving ICML 2025
Toward L_∞Recovery of Nonlinear Functions: A Polynomial Sample Complexity Bound for Gaussian Random Fields COLT 2023
Beyond NTK with Vanilla Gradient Descent: A Mean-Field Analysis of Neural Networks with Polynomial Width, Samples, and Time NIPS 2023
Model-Based Offline Reinforcement Learning with Local Misspecification AAAI 2023
First Steps Toward Understanding the Extrapolation of Nonlinear Models to Unseen Domains ICLR 2023
Asymptotic Instance-Optimal Algorithms for Interactive Decision Making ICLR 2023
Provable Model-based Nonlinear Bandit and Reinforcement Learning: Shelve Optimism, Embrace Virtual Curvature NIPS 2021
Design of Experiments for Stochastic Contextual Linear Bandits NIPS 2021
On the Expressivity of Neural Networks for Deep Reinforcement Learning ICML 2020
Root-n-Regret for Learning in Markov Decision Processes with Function Approximation and Low Bellman Rank COLT 2020
Multinomial Logit Bandit with Low Switching Cost ICML 2020
Q-learning with UCB Exploration is Sample Efficient for Infinite-Horizon MDP ICLR 2020
Exploration via Hindsight Goal Generation NIPS 2019